1. Which of the following correctly identifies a difference between cross-sectional data and time series data?

a. Cross-sectional data is based on temporal ordering, whereas time series data is not. b. Time series data is based on temporal ordering, whereas cross-sectional data is not.

c. Cross-sectional data consists of only qualitative variables, whereas time series data consists of only quantitative variables.

d. Time series data consists of only qualitative variables, whereas cross-sectional data does not include qualitative variables.

2. Refer to the following model

y t = α0 + β0st + β1st-1 + β2st-2 + β3st-3 + ut

This is an example of a(n):

a. infinite distributed lag model.

b. finite distributed lag model of order 1.

c. finite distributed lag model of order 2.

d. finite distributed lag model of order 3.

3. Which of the following is an assumption on which time series regression is based?

a. A time series process follows a model that is nonlinear in parameters.

b. In a time series process, no independent variable is a perfect linear combination of the others. c. In a time series process, at least one independent variable is a constant.

d. For each time period, the expected value of the error ut, given the explanatory variables for all time periods, is positive.

4. Adding a time trend can make an explanatory variable more significant if:

a. the dependent and independent variables have similar kinds of trends, but movement in the indepen variable about its trend line causes movement in the dependent variable away from its trend line.

b. the dependent and independent variables have similar kinds of trends and movement in the indepen variable about its trend line causes movement in the dependent variable towards its trend line.

c. the dependent and independent variables have different kinds of trends and movement in the indepe variable about its trend line causes movement in the dependent variable towards its trend line.

d. the dependent and independent variables have different kinds of trends, but movement in the indep variable about its trend line causes movement in the dependent variable away from its trend line.

5. Refer to the following model

y t = α0 + β0st + β1st-1 + β2st-2 + β3st-3 + ut

6. A stochastic process {xt: t = 1,2,….} is covariance stationary if:

a. E(xt) is variable, Var(xt) is variable, and for any t, h ≥ 1, Cov(xt, xt+h) depends only on ‘h ’ and not on ‘t’.

b. E(xt)  is variable, Var(xt) is variable, and for any t, h ≥ 1, Cov(xt, xt+h)  depends only on ‘t ’ and not on h.

c. E(xt) is constant, Var(xt) is constant, and for any t, h ≥ 1, Cov(xt, xt+h)  depends only on ‘h ’ and not on ‘t’.

d. E(xt)  is constant, Var(xt) is constant, and for any t, h ≥ 1, Cov(xt, xt+h)  depends only on ‘t ’ and not on ‘h ’.

7. The homoskedasticity assumption in time series regression suggests that the variance of the error term cannot be a function of time.

a. True

b. False

8. The HAC (robust) standard errors are typically larger than the usual OLS standard errors when there is serial correlation.

a. True

b. False

9.  The AR(p) model:

A) is defined as Yt = β0+ βpYt-p + ut.

B) represents Ytas a linear function of p of its lagged values.

C) can be represented as follows: Yt = β0+ β 1Xt+ βpYt-p + ut.

D) can be written as Yt = β0+ β 1Yt-1+ ut-p.

10. Departures from stationarity:

A) jeopardize forecasts and inference based on time series regression.

B) occur often in cross-sectional data.

C) can be made to have less severe consequences by using log-log specifications. D) cannot be fixed.

11. If the future differs fundamentally from the past, then:

A) the time series is stationary.

B) regression models estimated using past data can be used to forecast future values. C) historical relationships might not be reliable guides to the future.

D) the joint distribution of (, ,...  , does not depend on s, regardless of the value of T.

12. By including another variable in the regression, you will:

A) decrease the regression R2 if that variable is important.

B) eliminate the possibility of omitted variable bias from excluding that variable.

C) look at the t-statistic of the coefficient of that variable and include the variable only if the coefficient is statistically significant at the 1% level.

D) decrease the variance of the estimator of the coefficients of interest.

13. Errors-in-variables bias:

A) is only a problem in small samples.

B) arises from error in the measurement of the independent variable.

C) becomes larger as the variance in the explanatory variable increases relative to the error variance.

D) is particularly severe when the source is an error in the measurement of the dependent variable.

14. When testing joint hypothesis, you should:

A) use t-statistics for each hypothesis and reject the null hypothesis is all of the restrictions fail. B) use the F-statistic and reject all the hypothesis if the statistic exceeds the critical value.

C) use t-statistics for each hypothesis and reject the null hypothesis once the statistic exceeds the critical value for a single hypothesis.

D) use the F-statistics and reject at least one of the hypothesis if the statistic exceeds the critical value.

15. If you reject a joint null hypothesis using the F-test in a multiple hypothesis setting, then: A) a series of t-tests may or may not give you the same conclusion.

B) the regression is always significant.

C) all of the hypotheses are always simultaneously rejected.

D) the F-statistic must be negative.

16. All of the following are TRUE, with the exception of one condition:

A) a high R2 or  does not mean that the regressors are a true cause of the dependent variable. B) a high R2 or  does not mean that there is no omitted variable bias.

C) a high R2 or  always means that an added variable is statistically significant.

D) a high R2 or  does not necessarily mean that you have the most appropriate set of regressors.